{ "info": { "author": "Artem Streltsov", "author_email": "artem.streltsov@duke.edu", "bugtrack_url": null, "classifiers": [], "description": "C-SGHMC\r\n=======================\r\n\r\nThis is a repository created as part of a statistical computation project at Duke University.\r\n\r\nThe code to be found here is a C++ implementation of SGHMC algorithm wrapped for Python using pybind11. It is further used to verify the results in the original paper on SGHMC: \"Stochastic Gradient Hamiltonian Monte Carlo\" by Tianqi Chen, Emily B. Fox and Carlos Guestrin.\r\n\r\nAs seen from our project this is a **15-time-efficiency-improvement** in terms of runtime over the pure python version.\r\n\r\nMain files\r\n=======================\r\n- **sghmcwrap.cpp** is the original sghmc algo in C++ wrapped for Python\r\n\r\n- **c_sghmc.py** calls and decorates the above extension and tests on an example from the original paper \r\nPackages required to run the code\r\n=======================\r\n\r\n- The implementation makes use of **Eigen** library for C++ and hence run the following line:\r\n! git clone https://github.com/RLovelett/eigen.git\r\n\r\nMake sure you have it cloned in the same directory as your code.\r\n\r\n- **pybind11** and **cppimport** are the two standard packages required for wrapping C++ codes for Python and importing them:\r\n! pip3 install pybind11\r\n\r\n! pip3 install cppimport\r\n\r\nThe example makes use of **sympy** to specify functions (such as potential and kinetic energies as well as the gradient of the former), **numpy** and **numpy.random**.\r\n", "description_content_type": null, "docs_url": null, "download_url": "", "downloads": { "last_day": -1, "last_month": -1, "last_week": -1 }, "home_page": "https://github.com/astr93/c_sghmc", "keywords": "", "license": "", "maintainer": "", "maintainer_email": "", "name": "c-sghmc", "package_url": "https://pypi.org/project/c-sghmc/", "platform": "", "project_url": "https://pypi.org/project/c-sghmc/", "project_urls": { "Homepage": "https://github.com/astr93/c_sghmc" }, "release_url": "https://pypi.org/project/c-sghmc/0.0.1/", "requires_dist": null, "requires_python": "", "summary": "A C++ extension for SGHMC", "version": "0.0.1" }, "last_serial": 2843363, "releases": { "0.0.1": [ { "comment_text": "", "digests": { "md5": "b67eae8456785f987dc8f1d8bed7bc73", "sha256": "8afa9be6b3329be360ffa339d14f3d0997cc0c23996355f2a6d0d465acb26a9b" }, "downloads": -1, "filename": "c_sghmc-0.0.2.tar.gz", "has_sig": false, "md5_digest": "b67eae8456785f987dc8f1d8bed7bc73", "packagetype": "sdist", "python_version": "source", "requires_python": null, "size": 3006, "upload_time": "2017-05-01T21:31:39", "url": "https://files.pythonhosted.org/packages/6f/a7/9f60e3a0e7d7b08ae2b57673a47514673c6ea34c3f58909d8f1a29383508/c_sghmc-0.0.2.tar.gz" } ] }, "urls": [ { "comment_text": "", "digests": { "md5": "b67eae8456785f987dc8f1d8bed7bc73", "sha256": "8afa9be6b3329be360ffa339d14f3d0997cc0c23996355f2a6d0d465acb26a9b" }, "downloads": -1, "filename": "c_sghmc-0.0.2.tar.gz", "has_sig": false, "md5_digest": "b67eae8456785f987dc8f1d8bed7bc73", "packagetype": "sdist", "python_version": "source", "requires_python": null, "size": 3006, "upload_time": "2017-05-01T21:31:39", "url": "https://files.pythonhosted.org/packages/6f/a7/9f60e3a0e7d7b08ae2b57673a47514673c6ea34c3f58909d8f1a29383508/c_sghmc-0.0.2.tar.gz" } ] }